I don't use windows. In fact I've never really used windows, except occasionally surfing the web on someone else's machine.
I've always had Macs at home. At work I have used assorted Unix systems. This year I've gone completely Mac, replacing my Linux machine with a PowerMac G5.
Well, I went to a pretty good CS dept (UC Berkeley) and we did a heck of a lot of coding. For the compiler class, we wrote a compiler. For the OS class, we wrote multi-threaded, producer-consumer code. For the graphics class we wrote a software renderer.
You could avoid all that coding by doing more theory and hardware classes, but that would be pretty unusual.
Looking at IBM's accouncement it looks like they use same headers and run-time libraries as gcc 3.3. They say you can combine xlc and gcc compiled files. I.e., you can link object files.
According to this page a 64k processor CM-2 could do 2500 megaflops. Looking at the #4 machine on the Top 500 list, NCSA's P4 Xeon based system, a 3GHz Xeon gets about 3.9 gigaflops. But then it doesn't have cool blink lights of a CM-2.
Pretty amazing how far things have progressed. The first supercomputer, the Cray 1, introduced in 1976, did 160 megaflops and had 8 megabytes of memory. Kinda like a Palm Pilot.
Yeah, I'm with you. I worked in a lab with a CAVE for several years, and I've visited other places with CAVEs. They're mainly good for making gee-whiz demos to impress funding agencies. I have yet to see anyone actually doing real work in a CAVE.
I seriously considered doing a PhD at MIT but was put off by the 5 year course of which you're apparently treated like an undergraduate for the first 2 - compared to a 3 year course here.
I heard similar complaints from Europeans who came to the US to get PhDs. But I think from the American standpoint, if you get a doctorate in computer science (or whatever field) you should have some background in all the major areas of CS. If you just do research in some specialty (for me it was graphics), you would have missed taking classes in compilers, operating systems, architecture, etc.
Maybe in Europe they assume you've had that exposure in getting a masters degree. In America many programs don't require a CS masters to go for the PhD, so they have these course requirements. That's a big reason it's pretty rare to get a PhD in 3 years. My guess is the average is more like 6 years.
That's not necessarily true. Over at Stanford for the project they built a graphics system with 32 PCs that render to a tiled display. Imagine a display made of 1000 monitors in a 40x20 grid. That would be pretty freaking cool.
Really? What evidence do you have that the people on Taiwan want to be ruled by the folks that brought you the Tianamen Square massacre? Taken any opinion polls of that? All the Taiwanese people I know want to be able to decide for themselves who governs the island.
Pres. Bush has said that the US will defend Taiwan if China attacks. It's not so much about economics, but more about politics and being anti-communists. Even before Bush said this, it was implicit in the Taiwan Relations Act of 1979, but previous presidents let it unsaid.
Apparently James Gosling and a whole lot of developers at Sun are flocking to Macs. You'd think that would help spur on a native port of Open Office to Mac OS X.
This ZDNet story sayst that the 1989 suit got settled by Apple Computer paying Apple Records.
According to this article, Apple tried to get their insurance company to cover the $26.4 million settlement, but ultimately lost that decision in court.
Apple's web page says the throughput between the system controller and main memory is 2.7 Gigabytes per second. The problem is the CPU's are connected to the system controller on a 167 Megahertz, 64 bit MPX bus. That's 1.33 Gigabytes per second, and the two CPU's share the same bus. So the CPU's can't push data out (or suck it in) as fast as the memory can. Note that it's because of Motorola, not Apple.
On the other hand, the nice thing is that the graphics card and the disks can be accessing main memory at the same time as the CPU's and you won't have to worry about stalls because of contentions. This would probably be most evident on graphics intensive processing (3-d games) and database/web server type apps with lots of I/O.
Slashdot Mirror
I don't use windows. In fact I've never really used windows, except occasionally surfing the web on someone else's machine.
I've always had Macs at home. At work I have used assorted Unix systems. This year I've gone completely Mac, replacing my Linux machine with a PowerMac G5.
I've updated and it still works for me.
Well, I went to a pretty good CS dept (UC Berkeley) and we did a heck of a lot of coding. For the compiler class, we wrote a compiler. For the OS class, we wrote multi-threaded, producer-consumer code. For the graphics class we wrote a software renderer.
You could avoid all that coding by doing more theory and hardware classes, but that would be pretty unusual.
You missed some shit.
grep -EHirn "\Wshit\W" * | wc -l
40
And, if you put that 200 gig drive in your pocket, it's much more impressive to the ladies.
OK, how about a modern load/store architecture that isn't starved for registers?
Looking at IBM's accouncement it looks like they use same headers and run-time libraries as gcc 3.3. They say you can combine xlc and gcc compiled files. I.e., you can link object files.
According to this page a 64k processor CM-2 could do 2500 megaflops. Looking at the #4 machine on the Top 500 list, NCSA's P4 Xeon based system, a 3GHz Xeon gets about 3.9 gigaflops. But then it doesn't have cool blink lights of a CM-2. Pretty amazing how far things have progressed. The first supercomputer, the Cray 1, introduced in 1976, did 160 megaflops and had 8 megabytes of memory. Kinda like a Palm Pilot.
Wanna copy music off you iPod. Here's how I'd do it.
/Volumes/Dave Chen's iPod/iPod_Control/Music/* ~/Music
cp -r
If that's too complicated for ya, you can find apps that'll do it too.
For the newer iPods you can get a dual USB 2/Firewire cable. This Apple document shows how to charge and sync at the same time.
Yeah, I'm with you. I worked in a lab with a CAVE for several years, and I've visited other places with CAVEs. They're mainly good for making gee-whiz demos to impress funding agencies. I have yet to see anyone actually doing real work in a CAVE.
No, no, no. It's $2999 for the PowerMac, and $12,001 for the house repairs with it blows you away.
Maybe in Europe they assume you've had that exposure in getting a masters degree. In America many programs don't require a CS masters to go for the PhD, so they have these course requirements. That's a big reason it's pretty rare to get a PhD in 3 years. My guess is the average is more like 6 years.
That's not necessarily true. Over at Stanford for the project they built a graphics system with 32 PCs that render to a tiled display. Imagine a display made of 1000 monitors in a 40x20 grid. That would be pretty freaking cool.
Really? What evidence do you have that the people on Taiwan want to be ruled by the folks that brought you the Tianamen Square massacre? Taken any opinion polls of that? All the Taiwanese people I know want to be able to decide for themselves who governs the island.
Pres. Bush has said that the US will defend Taiwan if China attacks. It's not so much about economics, but more about politics and being anti-communists. Even before Bush said this, it was implicit in the Taiwan Relations Act of 1979, but previous presidents let it unsaid.
Apparently James Gosling and a whole lot of developers at Sun are flocking to Macs. You'd think that would help spur on a native port of Open Office to Mac OS X.
Yup, and OpenGL X11 programs use hardware rendering, unlike before.
Hey, if there's anyone that can find Bin Laden and bring him to justice, it's the BSA.
Ummm, MOAB, anyone?
This ZDNet story sayst that the 1989 suit got settled by Apple Computer paying Apple Records.
According to this article, Apple tried to get their insurance company to cover the $26.4 million settlement, but ultimately lost that decision in court.
Isn't that one of those Star Trek episodes? Kirk makes google search for itself and the computer blows up!
Apple's web page says the throughput between the system controller and main memory is 2.7 Gigabytes per second. The problem is the CPU's are connected to the system controller on a 167 Megahertz, 64 bit MPX bus. That's 1.33 Gigabytes per second, and the two CPU's share the same bus. So the CPU's can't push data out (or suck it in) as fast as the memory can. Note that it's because of Motorola, not Apple.
On the other hand, the nice thing is that the graphics card and the disks can be accessing main memory at the same time as the CPU's and you won't have to worry about stalls because of contentions. This would probably be most evident on graphics intensive processing (3-d games) and database/web server type apps with lots of I/O.
Furthermore, even if you could compile and run the OS, how would you know if trust the compiler? Maybe it's sticking a back door in the executable.
Well, I'm a Unix programmer. To get my code to work on a Mac, here's what I had to do:
make
There ya go.